diff --git a/README.md b/README.md
new file mode 100644
index 0000000..610b48d
--- /dev/null
+++ b/README.md
@@ -0,0 +1,45 @@
+# slm
+
+A simple command-line tool to interact with an LLM API.
+
+## Configuration
+
+The tool reads its configuration from `~/.config/slm/config.toml`.
+
+### Config File Format
+
+```toml
+api_key = "your_api_key_here"
+server_url = "https://your-api-server.com/v1/chat/completions"
+model = "nvidia/Gemma-4-31B-IT-NVFP4"
+```
+
+| Field | Description |
+| :--- | :--- |
+| `api_key` | Your API authentication key. |
+| `server_url` | The full URL to the chat completions endpoint. |
+| `model` | The identifier of the model to use. |
+
+## Usage
+
+By default, `slm` reads from standard input and sends it to the configured model.
+
+### Basic Example
+
+```bash
+echo "Hello, how are you?" | slm
+```
+
+### With a Prompt
+
+You can append a specific prompt to the input using the `-p` or `--prompt` argument. This is useful for giving the model a persona or specific instructions.
+
+```bash
+echo "the capital of France" | slm -p "Answer in one word."
+```
+
+## Arguments
+
+| Argument | Short | Description |
+| :--- | :--- | :--- |
+| `--prompt <TEXT>` | `-p` | Optional prompt to append to the content read from stdin. |
diff --git a/src/main.rs b/src/main.rs
index 9454ed9..3127ebe 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -15,6 +15,7 @@ struct Args {
 struct Config {
     api_key: String,
     server_url: String,
+    model: String,
 }
 
 #[derive(Serialize)]
@@ -70,7 +71,7 @@ async fn run() -> Result<(), Box<dyn std::error::Error>> {
 
     // Construct the API request
     let request_body = ChatRequest {
-        model: "nvidia/Gemma-4-31B-IT-NVFP4".to_string(), // Default model
+        model: config.model.clone(),
         messages: vec![Message {
             role: "user".to_string(),
             content: final_content,